I
I
L
P
E
E
D
R
I
I
Paris Interdisciplinary Energy Research Institute
Université de Paris
Ecole doctorale
Physique en Ile de France 564
Laboratoire Interdisciplinaire des Energies de Demain UMR 8236
Urban Materials and Evaporative Cooling for
Heat Mitigation in Cities
Adapting Pavement-Watering to Different Parisian
Pavements
Thèse de Doctorat
de Physique / Sciences pour l’Ingénieur
Dirigée par Laurent Royon
Réalisée par
Sophie Parison
Présentée et soutenue publiquement le 6 octobre 2020 Devant un jury composé de :
M. Stéphane Lassue Professeur, Université d’Artois Président du jury M. Mat Santamouris Professeur, University of New South Wales Rapporteur M. Valéry Masson Ingénieur en chef, CNRM Météo-France Rapporteur Mme Marjorie Musy Directrice de Recherche, Cerema Ouest Examinatrice M. Laurent Royon Professeur, Université de Paris Directeur
M. Martin Hendel Maître de conférences, ESIEE Paris Encadrant, Invité Mme Kristine Jurski Maître de conférences, Université de Paris Encadrante, Invitée Mme Agathe Cohen STEA, DPE, Mairie de Paris Invitée
This Ph.D. thesis was funded by the National Association for Research and Technology (ANRT) and Paris City Hall via the CIFRE programme. The research was carried out with the Paris In-terdisciplinary Energy Research Institute (Paris University) and with Paris City Hall, within the Technical Service for Water and Sanitation (Water and Sanitation Division), and the Public Space Laboratory (Road and Traffic Division).
Avant-propos
Cette thèse a été cofinancée par l’Association Nationale de la Recherche et de la Technologie (ANRT) et la Mairie de Paris par le biais du dispositif CIFRE (Conventions Industrielles de For-mation par la REcherche). Les travaux ont été effectués conjointement au sein du Laboratoire Interdisciplinaire des Energies de Demain (Université de Paris), ainsi qu’à la Ville de Paris, au Service Technique de l’Eau et de l’Assainissement (Direction de la Propreté et de l’Eau) et au Laboratoire de l’Espace Public (Direction de la Voirie et des Déplacements).
Abstract: This manuscript examines the use of pavement-watering as a heat mitigation strat-egy and climate change adaptation tool for cities. The method is fine-tuned for traditional and cool paving materials in order to limit the water footprint of the technique.
The first Part of this research is based on field measurements gathered form watering cam-paigns in Paris from 2013 to 2018. A suited statistical analysis method is proposed in order to determine the microclimatic effects of watering, including effects on air temperature and pedestrian thermal stress using the Universal Thermal Climate Index. Two watering protocols are compared to determine the influence of the surface area being watered on the efficiency and duration of pavement-watering.
Secondly, a laboratory experiment is used to compare the thermal behaviour of realistic paving structures under heat-wave like conditions. The pavement undergoes a 24-h climate cycle and watering can be enabled at a fixed frequency. On the basis of surface and in-depth temperature and heat flux measurements, using the surface heat budget, the evaporative cool-ing flux is determined for each tested watercool-ing rate. Results obtained on an asphalt road struc-ture with the lab protocol are compared to field results.
Finally, the lab protocol is applied to twelve traditional and cool pavements under dry and watered conditions. Watering is fine-tuned for each structure to maximize cooling and mini-mize the water consumption using two linear cooling regimes. Driving parameters influencing the optimization of the evaporative cooling versus the watering rate are determined. The sur-face heat budget and the partitioning of irradiance into conductive, convective, radiative and cooling fluxes are analysed for each paving structure. In the end, the benefits of each pavement, the efficiency of the method and the limitations of the lab protocol are discussed.
This research intends to provide useful information for decision-makers considering the use of pavement-watering or cool pavements as heat mitigation strategy. Future work should principally investigate the microclimatic effects of cool pavements combined with pavement-watering to confront lab results to field studies. Those should come with an adapted experi-mental design, while associated statistical procedures may also require improvements in the future.
Keywords:pavement-watering; urban materials; cool pavements; evaporative cooling; climate change adaptation.
Résumé : Ce manuscrit s’intéresse à l’arrosage urbain comme stratégie d’atténuation de la chaleur et d’adaptation au changement climatique. La méthode est optimisée pour des matéri-aux de voirie traditionnels et "frais" afin de limiter la consommation d’eau du procédé.
La première partie de ce manuscrit utilise des mesures de terrain recueillies lors de cam-pagnes d’arrosage à Paris entre 2013 et 2018. Une méthode d’analyse statistique permet de déterminer les effets microclimatiques de l’arrosage sur la température de l’air et le stress ther-mique à l’aide de l’Indice Universel de Climat Therther-mique. Deux protocoles d’arrosage sont comparés afin de déterminer l’influence de la surface arrosée sur l’efficacité de la méthode.
Dans un second temps, une expérience de laboratoire est utilisée afin de comparer le com-portement thermique de structures de voirie réalistes dans des conditions similaires à celles d’une vague de chaleur parisienne. L’échantillon subit un cycle climatique de 24 heures et son arrosage peut être activé à une fréquence fixe. A l’aide de mesures de température et de flux thermique en surface et en profondeur, le bilan thermique de surface permet de déterminer le flux rafraîchissant évaporatif pour chaque débit d’arrosage testé. Les résultats obtenus sur une structure de chaussée en asphalte sont comparés à ceux du terrain.
Enfin, le protocole est appliqué à douze revêtements traditionnels et "frais", avec et sans arrosage. L’arrosage est optimisé pour chaque structure afin de maximiser le rafraîchissement en minimisant la consommation d’eau. Pour ce faire, deux régimes de rafraîchissement sont employés. Les paramètres déterminants dans l’optimisation du flux rafraîchissant selon le débit sont identifiés. Le bilan thermique de la surface et le partitionnement de l’irradiance sont analysés pour chaque échantillon. L’efficacité de la méthode selon les structures ainsi que les limites du protocole de laboratoire sont discutées.
Ce travail vise à fournir des informations utiles aux décideurs qui envisagent l’emploi de l’arrosage urbain ou de revêtements "frais" dans leur stratégie d’atténuation de la chaleur. La recherche future devrait s’intéresser aux effets microclimatiques de revêtements "frais" en com-binaison avec de l’arrosage afin de confronter les résultats de laboratoire à davantage d’études de terrain. Ces dernières devront proposer un plan expérimental adéquat, tandis que les méth-odes d’analyse statistique associées pourront également faire l’objet de futures améliorations.
Mots-clés :arrosage urbain; matériaux urbains; revêtements frais; rafraîchissement évaporatif; adaptation au changement climatique.
Cette thèse constitue l’aboutissement de trois années de travail, riches d’apprentissages tant sur le plan scientifique que professionnel et humain. Je souhaiterais commencer ces quelques lignes en rappelant tout d’abord les circonstances dans lesquelles ce projet a vu le jour.
Initialement, à l’automne 2015, alors étudiante de Master 2, je n’envisageais pas encore pleinement la possibilité de faire une thèse. Presque par hasard, au fil de discussions passées avec Benjamin Thiria, responsable de mon Master, je compris que cette voie pourrait me plaire et me correspondre. C’est à l’occasion d’un projet de Master, portant sur les îlots de chaleur urbains et encadré par Martin Hendel et Laurent Royon, que l’opportunité de faire une thèse se présenta finalement. Ce projet m’aura permis de confirmer mon intérêt pour ces thématiques de recherche qui correspondaient déjà à mes sensibilités et mes attentes. Qui plus est, j’avais déjà eu l’opportunité de travailler avec Laurent Royon en stage de Master 1, ce qui me conforta dans mon envie de postuler à cette thèse au sein de l’équipe. La thèse proposée me permettait en outre d’intégrer la Ville de Paris par le biais du dispositif CIFRE, ce qui me satisfaisait com-plètement. Il eut fallu attendre décembre pour apprendre que ma candidature avait été retenue par Laurent et par Martin. Je tiens à les en remercier et à leur témoigner ma profonde gratitude pour la confiance qu’ils m’ont accordée. Puis, en avril 2016, vint la validation par le Conseil de Paris, et enfin en juin l’aval final et décisif de financement par l’ANRT, pour un début de thèse programmé en janvier 2017.
Je souhaite ici exprimer ma grande reconnaissance envers mon équipe d’encadrement de thèse. Je remercie en premier lieu mon directeur, Laurent Royon, pour son accompagnement tant scientifique qu’humain. Merci pour tes conseils et pour les discussions riches que nous avons eues, ainsi que pour tes commentaires pertinents sur mon travail. Je remercie également Martin Hendel, mon encadrant, pour m’avoir guidée et soutenue tout au long de cette thèse. Merci pour tes remarques, toujours très justes et constructives, et pour les très nombreuses réflexions communes que nous avons eues. Enfin, merci à vous deux pour votre bienveillance et votre grande prévenance à mon égard, dont la portée s’est même étendue au-delà du cadre de la thèse à diverses reprises (post-doc, enseignements, conférences, etc.). J’adresse aussi mes remerciements sincères à Agathe Cohen et Jérôme Lefebvre, mes encadrants Ville de Paris, pour leur aide logistique et pour leur écoute attentive, qui se sont avérées être essentielles à la réalisation pratique de mon travail de thèse. Merci également à Nicolas Londinsky et Damien Balland pour leur soutien managérial et opérationnel, ainsi qu’à Kristine Jurski pour son accompagnement tout au long de cette thèse. Merci enfin à Patricia Bordin et Brice Tréméac pour nos échanges et pour leur suivi au sein de mon comité de thèse.
Je remercie mes rapporteurs Mat Santamouris et Valéry Masson, ainsi que mes examina-teurs, Marjorie Musy et Stéphane Lassue, pour avoir accepté de faire partie de mon jury, pour la relecture de mon manuscrit ainsi que pour leurs remarques constructives et les échanges que cela a nourri.
Je souhaite aussi adresser mes remerciements aux services de la Ville de Paris avec lesquels j’ai collaboré durant ma thèse. Merci aux équipes du STEA, en particulier à la SPE pour son ac-cueil, mais également à la SAP, la DEI, la DII et au STPP. Je remercie en particulier les équipes iv
opérationnelles ayant mis en œuvre l’arrosage l’été pour me permettre de mener ce projet à bien, ainsi que l’atelier Deleusseux pour son aide dans la pose et dépose de stations météo sur le terrain. Merci aux collègues de Commandeur pour les bons moments passés ensem-ble : Fenotte, Nathalie, Marlène, Michelle pour son accueil au 5eme, Khajin, Anastasia, Chantal, Marie-Claude, Daouda, Alain et Jean-Claude. Je salue aussi les membres du groupe de travail Expérimentation de rafraîchissement de l’espace public par arrosage que j’ai piloté durant ma thèse, mais également ceux du groupe Ilots et parcours de fraîcheur piloté par l’AEU, auquel j’ai eu le plaisir de contribuer. J’adresse également mes remerciements au LEP, qui m’a aussi accueil-lie durant cette thèse. Merci aux équipes opérationnelles pour leur accueil chaleureux, leur disponibilité et pour leur aide pratique, notamment dans la confection d’éprouvettes, essen-tielles à la réalisation du travail décrit dans ce document. Un grand merci à Didier, Isabelle, Khelil, Eric, Souazic, Jean-Luc, Céline, Laurent, Claude, Chantal et Patricia. Enfin, merci à So-phie Zuber de la DAE pour son écoute attentive et son accompagnement précieux pendant ces trois années.
Je tiens également à remercier les personnes avec qui j’ai eu l’occasion de collaborer à di-verses reprises durant ces trois années. Je remercie tous les membres du projet Life Cool & Low Noise Asphalt auquel j’ai eu la chance de contribuer : merci à la DEVE et à l’AEU, à nouveau à la DPE et à la DVD, mais également à Colas, Eurovia et BruitParif. Je salue également Andrej et Marcos ainsi que tous les partenaires du projet Lisière d’une Tierce Forêt. Merci à SLG et à Captec pour leur aide précieuse concernant les instruments de mesure employés pendant ma thèse. Merci également aux membres du Workshop Climat Urbain pour les échanges que nous avons pu avoir. J’exprime enfin ma gratitude envers Valérie Douay, Emmanuelle Algré et Gaëlle Lissorgues pour leur chaleureux accueil à l’ESIEE à l’occasion de vacations d’enseignement, ainsi qu’à Kristine, à nouveau, en ce qui concerne les vacations à l’IUT de Paris-Diderot. Merci également aux stagiaires que j’ai encadrés et qui m’ont apporté une aide précieuse dans mon travail durant l’été : Eva, Param, et Maïlys, mais également Khadija, Lucas O, Lucas B, André et Antoine.
Je remercie aussi les anciens et actuels membres du LIED et du MSC ainsi que les techni-ciens, pour leur aide, pour nos discussions et pour le temps passé ensemble : Fred, José, Petros, Catherine, Xiao-Feng, Delphine, Insaaf, Melina, Wlad, Amir, Mathieu, Christophe, Eric, Syl-vain, Cécilia, Florence, Gwenaëlle, et tous les autres. Un très grand merci à Arnaud pour son aide considérable dans la réalisation de l’expérimentation de laboratoire, ainsi que pour les bons moments passés ensemble. Je n’oublie pas les membres de la 544 A, Etienne, Sara, Nico-las et Darius, et remercie aussi Grégoire du MPQ, ainsi que Marion, Yann et Donatella du côté de l’APC.
Je souhaiterais faire part de mes sincères remerciements aux doctorants du LIED et du MSC aux côtés desquels j’ai eu la chance de réaliser ma thèse depuis son commencement (ou presque) : Marie-Cécile, Rémy, Souhil, Yohann, Joachim, Zejun et Chuanyu. Aussi, c’est avec beaucoup d’affection que je voudrais tout particulièrement remercier les doctorants de l’équipe CEMU. Au début de ma thèse, j’ai intégré cette équipe en tant que seule doctorante travaillant sur le rafraîchissement urbain. Il m’aura fallu patienter jusqu’à l’été 2018 avant de voir arriver des stagiaires au sein de l’équipe sur cette même thématique, qui comme moi se seront finale-ment transformés en "stagiaires longue durée", avant de progressivefinale-ment devenir des amis.
Merci à vous, Maxime, Ghid, Maïlys et Maxime F, pour nos éclats de rire, nos discussions an-imées, nos pauses café, nos crises existentielles, nos instants "mezza", nos virées en triporteur (non sans risques !), nos verres après le travail, et pour tout le reste. Je t’adresse aussi un merci particulier Marie-Cécile, amie et "compagne de thèse" dès le début de celle-ci (bien que sur des sujets très différents), et ce jusqu’à nos soutenances respectives qui ont eu lieu à deux jours d’intervalle. J’en profite pour remercier Martin une fois encore, pour son initiation à la pré-paration et à la dégustation de café "digne de ce nom", ainsi que pour ses (très) nombreuses relectures indispensables de ce manuscrit. Merci à toute l’équipe pour sa préparation de la sou-tenance et son organisation du buffet qui a suivi, mais surtout, merci à vous tous pour votre bonne humeur quotidienne, pour tout ce que vous m’avez apporté humainement, et pour nos moments partagés, qui je l’espère perdureront bien au-delà du cadre de nos thèses.
Je voudrais aussi remercier mes amis pour leur soutien plus que nécessaire : Amina, Aman-dine, Anne-So, Céline, Julien, Gab, Thomas, Raph, Marion, Emma et tous les autres. Je remercie aussi toute ma famille pour sa présence et son soutien. Merci en particulier à mes deux sœurs adorées Marie et Claire, ainsi qu’à ma mère, à qui je dois pour beaucoup mon parcours uni-versitaire. Merci également à toi Philippe, pour ta présence quotidienne à mes côtés depuis maintenant plusieurs années, dans les moments ordinaires comme dans les moments difficiles, mais aussi pour ton soutien scientifique et moral et pour la stabilité affective que tu continues de m’apporter chaque jour. Merci enfin à toutes les personnes que j’oublie et qui ont contribué de près ou de loin à la réussite de cette thèse.
Contents vii
List of Figures xi
List of Tables xiii
Acronyms and Abbreviations xv
List of Symbols xvii
I GENERAL INTRODUCTION 1
1 Introduction 3
2 Scientific Background 7
2.1 The UHI Effect and the Urban Climate . . . 7
2.2 Urban Heat Budget . . . 9
2.3 Pedestrian Heat Stress . . . 12
2.3.1 Radiative Environment and Mean Radiant Temperature . . . 13
2.3.2 Thermal Stress Indices . . . 15
2.3.3 Cooling Strategies and the Human Heat Budget . . . 18
2.4 Urban Heat Mitigation Strategies. . . 19
2.4.1 Cool Materials. . . 19
2.4.2 Urban Greening and Shading . . . 24
2.4.3 Other Cooling Techniques . . . 24
2.5 Conclusion . . . 25
3 Literature Review of Pavement-Watering 27 3.1 Methodology . . . 29
3.1.1 Scale and Approach . . . 29
3.1.2 Watering Strategy and Target Area . . . 30
3.1.3 Protocol Optimization and Water Consumption . . . 33
3.1.4 Dry-Wet Difference Assessment Method. . . 34
3.2 Reported Cooling Effects of Pavement-Watering . . . 36
3.2.1 Microclimatic Effects . . . 37
3.2.2 Thermal Effects . . . 40
3.3 Conclusion . . . 43
4 Research Questions Adressed 47 II FIELDEFFECTSASSESSMENT OFUHI-MITIGATIONTECHNIQUES 49 5 Introduction to Part II 51 6 Methodology 53 6.1 Mathematical Framework . . . 53
6.2 Suited Statistical Tests . . . 55
6.3 Pavement-Watering . . . 58
6.3.1 Watering Protocol . . . 58
6.3.2 Sites Characteristics and Instrumentation . . . 58
6.3.3 Watering Criteria . . . 60
7 Results 63 7.1 Effects of Street and Sidewalk Watering: 2013 to 2015 . . . 63
7.2 Effects of Street-Only Watering: 2016 to 2018 . . . 66
8 Discussion of Part II 69 9 Conclusion of Part II 71 III LABORATORYOPTIMIZATION OF PAVEMENT-WATERING 73 10 Introduction to Part III 75 11 Methodology 77 11.1 Experimental Set-Up and Protocol . . . 77
11.2 Case-Study Structure . . . 80
11.3 Heat Transfer Analysis . . . 81
11.4 Watering Optimization Goals . . . 84
12 Experimental Results 85 12.1 Surface Temperature . . . 85
12.2 In-Depth Temperature . . . 87
12.3 In-Depth Conductive Heat Flux . . . 88
12.4 Atmospheric Convective Heat Flux . . . 90
12.5 Stored and Released Energy . . . 91
12.6 Evaporative Cooling Flux . . . 92
13 Discussion of Part III 95
IV HEAT BUDGETS OF STANDARD, COOL ANDWATEREDPAVEMENTS 101
15 Introduction to Part IV 103
16 Methodology 105
16.1 Studied Pavement Structures . . . 105
16.2 Albedo and Emissivity . . . 107
16.3 Thermal Conductivity . . . 108
16.4 Watering Optimization Goals . . . 109
17 Pavement Dry Behaviour 111 17.1 Surface Temperature . . . 111
17.2 In-depth Temperature . . . 113
18 Optimization of Watering 115 18.1 Temperature Reductions With Watering . . . 115
18.2 Conductive, Convective and Radiative Dry-Wet Variations . . . 117
18.3 Total Pavement Cooling Flux . . . 118
19 Steady-State Thermal Analysis 123 19.1 Surface Partitioning of Irradiance. . . 123
19.2 Total Cooling Flux Constitution. . . 125
19.3 In-Depth Heat Transmission . . . 129
20 Discussion: Transferability to the Field 131 21 Conclusion of Part IV 133 V GENERAL CONCLUSION 137 22 Conclusion 139 22.1 Field Testing . . . 139
22.1.1 Analysis Method . . . 140
22.1.2 Microclimatic Effects and Proportion of Street Watered . . . 140
22.2 Laboratory Testing . . . 141
22.2.1 Optimization of Watering . . . 141
22.2.2 Energy Partitioning of Dry and Optimally Watered Pavements. . . 142
23 Discussion and Future Research 145
Bibliography 152
Appendices 167
A A Radiative Technique for Measuring the Thermal Properties of Road and
B 5thInternational Conference on Countermeasures to Urban Heat Islands 187
C 33èmeColloque Annuel de l’Association Internationale de Climatologie 197
2.1 Urban surface heat budget. . . 9
2.2 Urban volume heat budget. . . 11
2.3 Human heat budget. . . 13
2.4 Energy balance of a globe thermometer. . . 14
3.1 Illustration of the heat budget of a dry and watered surface. . . 27
6.1 BACI design: control and case sites, before and after countermeasure imple-mentation. . . 53
6.2 Conceptual representation of a LMM applied to our framework. . . 57
6.3 Illustration of the case and control portions in rue du Louvre and photograph of a cleaning truck performing watering. . . 59
6.4 Photographs of the case and control weather stations in rue du Louvre. . . 60
7.1 Average watering effects at Louvre from 2013 to 2015. . . 64
7.2 Average watering effects at Louvre from 2016 to 2018. . . 67
11.1 Diagram and photograph of the experimental set-up. . . 77
11.2 Temperature and relative humidity regulation over 24 h inside the climate cham-ber. . . 78
11.3 Spectral irradiance of the halogen lamps and of the AM 1.5 global horizontal solar spectrum in the 200−1700 nm band . . . 79
11.4 Asphalt-road-sample diagram and photograph. . . 80
11.5 Schematic diagram of heat balance at the pavement surface. . . 81
12.1 Asphalt road surface temperature for various watering rates. . . 85
12.2 Asphalt road surface temperature during day phase with and without watering for 0.1 and 0.75 mm/h.. . . 86
12.3 Dry-wet maximum surface temperature difference for lower and upper fronts as a function of the watering rate for the asphalt road. . . 87
12.4 In-depth asphalt road temperature for non-watered and watered trials. . . 88
12.5 In-depth dry-wet maximum temperature difference as a function of the water-ing rate for the asphalt road. . . 88
12.6 Asphalt road in-depth conduction heat flux for dry and watered trials. . . 89
12.7 Average day-phase dry-wet difference in conduction heat flux to watering rate, for 6 and 14 cm deep asphalt road signals. . . 89
12.8 Asphalt road convective heat flux during day and night phases versus surface-air temperature difference. . . 90
12.9 Stored (day phase) and released (night phase) pavement surface conduction energy density, and released (day and night phases) atmospheric convective
energy density, to watering rate. . . 91
12.10 Dry-wet differences in net radiation, absorbed surface heat flux and atmospheric convective heat flux, and total pavement cooling flux to watering rate at the end
of day phase. . . 92
12.11 Latent and sensible cooling flux to watering rate. . . 93
12.12 Evaporation rate of the water-film sprinkled during the day phase as a function
of the watering rate. . . 94
16.1 Pavement structures studied in the lab. . . 105
16.2 Photograph of some of the tested samples. . . 106
16.3 Spectral reflectance averaged over several trials for each dry pavement structure.107
16.4 Schematic representation of the so-called latent and sensible cooling regimes
when neglecting sensible cooling for each regime. . . 110
17.1 Surface temperatures of standard and innovative pavement structures over a
24-h non-watered trial.. . . 111
17.2 Daily surface temperature increase during a dry test versus albedo and the
ab-sorptivity index. . . 112
17.3 In-depth temperatures of standard and innovative dry pavement structures 6,
14 and 25 cm deep. . . 114
18.1 Maximum surface temperature dry-wet difference versus watering rate for
tra-ditional and innovative pavements. . . 115
18.2 Dry-wet differences of conductive heat flux, net radiation and convective flux,
versus watering rate, for standard and innovative pavements. . . 117
18.3 Total cooling flux as a function of the watering rate for traditional and
alterna-tive pavements. . . 119
18.4 Optimal watering rate for each pavement structure as a function of their albedo
and their absorptivity. . . 121
19.1 Steady-state surface partitioning of irradiance into radiosity, conduction heat flux, atmospheric convective heat flux and cooling flux, all divided by
irradi-ance and plotted as a function of the absorptivity index a for all pavements. . . . 124
19.2 Detailed constitution of steady-state cooling flux normalized by irradiance for
optimally watered trials.. . . 126
19.3 Steady-state heat fluxes in-depth to irradiance as a function of the transmission
index for all pavements during a dry trial and an optimally-watered trial. . . 129
23.1 Steady-state net radiation to irradiance as a function of the absorptivity for
non-watered pavements. . . 147
23.2 Daily net radiation at Louvre for the case site from summers 2014 to 2018. . . 148
3.1 Experimental, numerical or combined articles among the literature studying
pavement-watering at slab-, street-, district- or city-scale.. . . 29
3.2 Detailed pavement-watering methods used in the surveyed studies. . . 31
3.3 Reported maximum air temperature effects. . . 38
3.4 Reported maximum air humidity effects. . . 38
3.5 Reported maximum effects on the radiative environment. . . 38
3.6 Reported maximum UHI-mitigation effects. . . 39
3.7 Reported maximum thermal comfort effects. . . 39
3.8 Reported maximum surface temperature effects.. . . 40
3.9 Reported maximum pavement temperature effects. . . 41
3.10 Reported maximum latent heat flow densities. . . 42
3.11 Reported maximum pavement heat flow densities. . . 43
3.12 Literature review summary for maximum dry-wet differences on ground sur-faces only. . . 45
6.1 Watering protocols carried out from 2013 to 2015 and from 2016 to 2018. . . 58
6.2 Instrument type, measurement height and uncertainty. . . 59
6.3 Weather conditions required for pavement-watering and heat-wave warnings. . 60
7.1 Duration, average and maximum values and occurrence hour of maximum ef-fect for statistically significant events of pavement-watering at Louvre between 2013 and 2015 using a linear fixed-effects model.. . . 65
7.2 Duration, average and maximum values and occurrence hour of maximum ef-fect for statistically significant events of pavement-watering at Louvre between 2016 and 2018 using a linear fixed-effects model.. . . 66
8.1 p-value and associated 24-hour average stat. sign. effect at Louvre for the 2013−2015 and 2016−2018 campaigns. . . 69
11.1 Characteristics of the day and night phases. . . 78
11.2 Watering rates tested. . . 79
11.3 Typical radiative balance of a mid-latitude city of about one million inhabitants under clear summertime skies and low wind speeds. . . 80
13.1 Confrontation of lab and field observations for optimal watering rates found in both cases. . . 96
16.1 In-depth position of the thermal sensors for each structure. . . 106
16.2 Albedo and emissivity of the studied paving structures. . . 108
16.3 Apparent thermal conductivity of pavement structures’ upper layers. . . 108
18.1 Best fitting equations for both cooling regimes and corresponding optimal
wa-tering rate for each pavement. . . 116
18.2 Optimal value of Φ, best fitting equations for both regimes, optimal watering
rate Qopti determined with the piecewise regression, and expected optimal rate
neglecting sensible cooling, for all pavements. . . 120
19.1 Cumulative absorbed daytime radiation of dry pavements and detailed
BACI Before-After-Control-Impact com. combined
counter countermeasure period (after countermeasure implementation) exp. experimental
FEM fixed-effects model
GIS geographic information system
HD Humidex
HIP heat island potential
IR infrared
LCZ local climate zone LMM linear mixed model LW longwave (3 - 100 µm)
MBACI multiple Before-After-Control-Impact design MCTP mass convection transport problem
MEMI Munich Energy Balance Model for Individuals MRT mean radiant temperature
NA not available
NIR near-infrared (0.74 - 3 µm) NR not relevant
num. numerical
PCM phase-change material
PET Physiological Equivalent Temperature PMV Predicted Mean Vote
PPCC pervious Portland-cement concrete
ref reference period (before countermeasure implementation)
RH relative humidity
SET* Standard Equivalent Temperature stat. sign. statistically significant
SW shortwave (0.3 - 3 µm) TEB Town Energy Balance UHI urban heat island
UTCI Universal Thermal Climate Index WBGT Wet-Bulb Globe Temperature WCI Wind Chill Index
Aitx local landscape effects for a weather type i, period t and location x,[−]
α albedo or shortwave reflectivity,[−]
a global surface absorptivity,[−]
Citx "background" climate for a weather type i, period t and location x,[−]
c specific heat,[J.kg−1.K−1]
cw specific heat of water: 4.18 kJ.kg−1.K−1
ca specific heat of air: 1.005 kJ.kg−1.K−1
∆Hdry−wet dry-wet difference in atmospheric convective heat flux,[W.m−2]
∆Mt interstation profile for the period t,[−]
∆Q sensible heat absorption flux,[W.m−2] ∆Rdry−wet
n dry-wet difference in net radiation,[W.m−2]
∆T temperature reduction,[°C] ∆Vdry−wet
0 dry-wet difference in surface conductive heat flux,[W.m−2]
E evaporation rate,[mm/h]
Emax maximum evaporation rate,[mm/h]
E0 surface conduction energy density,[kWh.m−2]
EH, atm atmospheric convection energy density,[kWh.m−2]
e surface material emissivity,[−] et deviation from the linear model,[−]
ew emissivity of water: 0.98
e thickness,[m]
H upwards atmospheric convective flux,[W.m−2] H/W canyon aspect ratio,[−]
h convective heat transfer coefficient,[W.m−2.K−1] Ix countermeasure impact at location x,[−]
i weather type,[−]
K total number of datasets,[−]
k k-th dataset among the total number of datasets,[−] L incident longwave radiation,[W.m−2]
Lup upwards longwave radiation,[W.m−2]
λ thermal conductivity,[W.m−1.K−1]
l latent heat of vaporization of water: 2,260 kJ.kg−1
Mitx meteorological parameter for a weather type i, period t and location x,[−]
Nt total number of observations for a given time period t,[−]
Φ total pavement cooling flux,[W.m−2] Φlat latent cooling flux,[W.m−2]
Φopti optimal pavement cooling flux,[W.m−2]
Φreg, lat "latent" regime cooling flux,[W.m−2]
Φreg, sens "sensible" regime cooling flux,[W.m−2]
Φsens sensible cooling flux,[W.m−2]
ϕg global surface heat flow,[W.m−2]
ϕrad radiative surface heat flow,[W.m−2]
p0 total air pressure,[Pa]
ps saturation vapour pressure,[Pa]
pv partial vapour pressure,[Pa]
Q watering rate,[mm/h] QA heat advection,[W.m−2]
QF atmospheric anthropogenic heat release,[W.m−2]
Qopti optimal watering rate,[mm/h]
Rintercept random intercept,[−]
Rn net radiation,[W.m−2]
Rslope random slope,[−]
ρ density,[kg.m−3]
ρw water density: 1000 kg.m−3
S incident shortwave radiation,[W.m−2] Sup reflected shortwave radiation,[W.m−2]
σ Stefan-Boltzmann constant: 5.67×10−8W.m−2.K−4
Ta air temperature,[°C]
Tdew dew point temperature,[°C]
Tg globe temperature,[°C]
Tmrt mean radiant temperature,[°C]
Ts surface temperature,[°C]
Tw water temperature,[°C]
Twet−bulb wet-bulb air temperature,[°C]
Tz pavement temperature at depth z,[°C]
τ in-depth heat transmission index,[−]
t period of evaluation (before or after countermeasure implementation,[−], or time,[s] Uitx effects of local urbanization for a weather type i, period t and location x,[−]
V downwards conductive heat flux,[W.m−2]
Vz downwards conductive heat flux at depth z[cm],[W.m−2]
v wind speed,[m.s−1]
x1 control site location,[−]
x2 case site location,[−]
G
ENERAL
I
NTRODUCTION
Introduction
The strong cooling benefits of water under the overwhelming heat of hot climates is a well-known fact. Arabic architecture took full advantage of it using mashrabiyas, decorative win-dows favouring ventilation, behind which porous clay jars filled with water were stored. Water evaporation cooled the passing air as it entered the house (Mohamed,2015). Similarly, wealthy ancient Persian homes sometimes came with special rooms connected to badgirs, wind-catching towers that forced downwards air circulation, where cooling was often enhanced with a small water pond or an underground water channel called qanat (English,1998). These provided a cool refuge during the afternoon in the harsh Iranian climate.
Examples of such vernacular designs or historical practices can be found abundantly all across the globe. In traditional Japan, the secular water-throwing ceremony of uchimizu (from uchi, to throw and mizu, water) plays a major cultural role in addition to cooling ground sur-faces and the local atmosphere during summers. As well, ground- or roof-sprinkling with water has been employed for centuries as a customary practice in the Mediterranean region, although it is not limited to hot areas. For instance, in his novel Au bonheur des dames published in 1883, Emile Zola mentions its informal practice in Paris: "Un lourd soleil chauffait les vitrages, et malgré les stores de toile grise, la chaleur tombait dans l’air immobile. Par moments, une haleine fraîche montait des parquets, que des garçons de magasin arrosaient d’un mince filet d’eau1."
In Paris still, official traces of this practice by the municipal services can be found in the early 20th century. At that time, public spaces were regularly watered in order to placate dust clouds formed by passing horse-drawn carriages. If the modernization of transport has caused the gradual disappearance of this practice up to this day, local inhabitants of the time reported a sensation of cooling associated with watering (Girard,1923). Street-cleaning with non-potable water was enabled thanks to the progressive modernization of the water supply and drainage infrastructures of the French capital. The latter is now a far cry from the Gallic Lutetia, which used to draw its water directly from the river, shortly before the first construction of the Roman aqueducts. From 1860, along with the important renovation of Paris driven by Baron
Hauss-1"A warm sun was playing on the windows, and in spite of the grey linen blinds, the heat penetrated the stagnant air. Now
and then a refreshing breath arose from the floor, which some assistants were gently watering." (translated from The Ladies’ Paradise, Emile Zola, 1883)
mann and in accordance with the hygienist theories, the engineer Eugène Belgrand supervised the creation of the underground sewers, within which two twin water networks, potable and non-potable, were installed (Husson,1996). Non-potable water, principally sourced from the Ourcq canal and to a lesser extent from the Seine, remains dedicated to public use today, sup-plying public fountains, parks and street-cleaning.
In the last couple of decades, pavement-watering has regained popularity across the globe, with scientific field studies and publications principally in Japan or France. Specifically, in Paris, the use of the non-potable water network for emergency cooling of public spaces during the summer has recently been of particular interest for the City Hall. Indeed, following a deci-sion by the City Council in 2012, the non-potable water network has been given new impetus
(Paris City Council, 2012). Ever since, it has been undergoing renovation work as well as a
revaluation of its uses. In this context, its use to cool the city has been studied notably as part of a Ph.D. thesis which aimed to quantify the cooling effects of this technique (Hendel,2015).
This initiative is part of a wider political movement driven by the City of Paris for several years in order to develop a wide range of climate change adaptation measures. This awareness rose in the wake of the major European heat-wave of 2003, shortly followed by Paris’ first Cli-mate Plan in 2007. Since then, the reinforcement of these measures is also motivated by predic-tions of an increase in the frequency, duration and intensity of heat-waves as a result of climate change in the years to come (Lemonsu et al.,2013). Awareness has been further strengthened by recent weather observations, which indicate that 2019, 2018 and 2014 were the hottest years on record in France (Météo-France, 2020). While the Climate Plan is updated every 5 years
(Paris City Hall, 2017a), Adaptation and Resilience strategies have also been adopted in the
meantime (Paris City Hall,2015,2017b). These milestones have laid a favourable ground for the emergence of various studies (Météo-France and CSTB,2012) and urban planning projects all aiming at cooling public spaces via urban greening, reflective materials or new water uses such as pavement-watering.
In this context, cities are seeking guidance in the implementation of their cooling strategies at local territorial scale, taking into account the specificities of each site and using suited evalu-ation tools. The main goal of this dissertevalu-ation is to provide elements to advise decision-makers on the procedure to favour, with a specific focus on pavement-watering and pavement materi-als. Many facets can be considered to tackle the problem. For instance, water availability, posi-tive impacts on air pollution or soil cleanliness and other potential socio-economic side-effects are relevant when deploying a pavement-watering strategy and should not be overlooked. The use of appropriate paving structures, including cool materials, raises questions about mechani-cal resistance, adhesion, maintenance, ageing, reparability, availability or implementation pro-cedures. Regardless of the solution considered, associated energy and financial costs as well as CO2 emissions are also important. In order to narrow the area of research, we turn to the
language associated with adaptation to heat-waves in Paris’ Climate Plan (2017a). This brings us to focus on the extensive study of the microclimatic and thermal effects of cooling strategies through the prism of public-space intervention. The document prompts the development of specific areas "open to the public and cooler than the nearby environment" and solutions that help "reduce the temperature". Having defined these criteria, we first identify decision-makers’ specific needs.
Practically speaking, whether for pavement-watering or other cooling strategies, the effi-ciency of the implemented solution needs to be quantitatively assessed. In order to do so, adapted tools and methods for in situ evaluation of their performance under real conditions must be developed. The field deployment of a pavement-watering strategy also requires know-ing which surface area (whole street, pavement, sidewalk, etc.) needs to be treated in order to obtain satisfactory cooling benefits while limiting the method’s water consumption. Assess-ing the coolAssess-ing effects associated with a given treated street-portion also meets an operational need, as the potential presence of shops, terraces or pedestrians on Parisian sidewalks can pre-vent from implementing watering in a simple manner. Decision-makers therefore need to be aware of the performance associated with each possible strategy.
Finally, declining pavement-watering to different sites requires adapting the strategy to their specificities (type of materials in place, exposure, etc.), as those influence the optimal wa-tering strategy to adopt in order to reduce the water consumption associated with pavement-watering. At the same time, urban-planning projects integrating innovative materials are emerging, as alternatives to traditional asphalt pavements inherited from the post-Haussmanian period. To answer this question without deploying costly and long-lasting field tests, a labora-tory approach can be used, while validated against a limited number of field studies. Such an approach permits optimizing the watering strategy according to the pavement material, and studying the behaviour of each pavement material and its synergies with watering.
Our research question can therefore be stated as follows: how can a pavement-watering strategy be operationally adapted to the characteristics of different urban sites, in particular to the variety of paving materials composing the urban fabric and to the surface area available for watering? With which tools can the corresponding microclimatic effects be evaluated?
Before getting to the very heart of the matter, the following Chapters of PartIwill present the scientific background of urban climate, thermal comfort and heat mitigation strategies. A brief state of the art of pavement-watering will be provided as well. These Chapters will help us identify more specific research questions that will guide this dissertation.
Scientific Background
2.1
The UHI Effect and the Urban Climate
Urban areas, through a combination of various mechanisms, display warmer air and surface temperatures than their surroundings. This phenomenon, known as the urban heat island (UHI) effect, is a tangible expression of the presence of cities and the human activities taking place within them. The UHIeffect has long been observed in many cities around the globe. It was first described in the 1810’s by Luke Howard, commonly referred to as the father of modern meteorology, who established a very thorough documentation of London’s climate
(Howard,1833). Through temperature records outside and within London, Howard observed
that "the temperature of the city is not to be considered as that of the climate: it partakes too much of an artificial warmth, induced by its structure, by a crowded population and the consumption of great quantities of fuel in fires". Emilien Renou, a French meteorologist, conducted a similar work for Paris during the mid-19th century (Renou, 1855). The phenomenon was named later on by
analogy to the topographic patterns of the isotherms on a surface weather map, which appear as an island (Stull,2012).
TheUHIphenomenon is the most well-known feature of the urban climate. Lowry(1977) was the first to establish a framework choosing a site’s past-atmospheric state as reference for explaining the "urban effect". Urban meteorological observations are seen as the "background climate" (the "flat-plane" climatological background for the region) on which the "local climate" (function of local natural features such as topography, shorelines, etc.) as well as "urbanization effects" are overlaid. The understanding of the physical mechanisms involved in the urban heat budget has since been significantly refined. In short,UHIsare caused by several factors. Urban morphology, because of wind and sky obstruction, reduces convective exchanges in ad-dition to favouring radiative trapping. During the day, solar radiation tends to be absorbed by urban surfaces through multiple reflections, as does the infrared radiation emitted by mineral surfaces at night. The higher-inertia urban fabric also favours heat storage, while the lack of vegetation reduces latent exchanges and anthropogenic heat release further increases city-scale heat gains (Oke,1982). Various studies have examined the link between some of these mech-anisms and urban configurations, such as the relationship between the sky view factor (form 7
factor between a point at ground level and the sky, dependent on sky masks such as trees, buildings, etc.) andUHIintensity, or that between radiative trapping and the canyon aspect ratio (H/W) (Oke,1981,1988a).
TheUHIintensity (or magnitude), measured as a simultaneous temperature difference be-tween an urban area and its surroundings, spans on average bebe-tween 1° and 3°C (Grimmond,
2007). For example, Paris’ urban core is 3°C warmer on a year average than its rural surround-ings (Cantat,2004). However,UHIis both spatially and temporally dependent, in addition to being related to a city’s intrinsic characteristics (e.g. size, shape, building density, land-use, population, etc.) and external influence (local climate, seasons, weather). Because of the differ-ent behaviour of the city and its surroundings,UHIintensity can be negative in the morning (so-called urban cool island) when rural areas with a lower thermal inertia warm up faster than the city as the latter achieves its heat release. On the contrary, its maximum intensity is reached before sunrise (Oke,1982).
Under clear skies and in the absence of wind, this difference can reach up to 10°C ( Grim-mond,2007). This is typically the case during heat-waves, which are worsened byUHI, which in turn contributes to worsen their health impacts, and has been observed in Paris (Cantat,
2004; Météo-France and CSTB, 2012) as for other cities around the world (Li and Bou-Zeid,
2013). Thus, in addition to altering the local atmosphere, heat islands affect human thermal comfort and public-health, by increasing the heat-related morbidity (Grimmond, 2007). For instance, the 2003 European heat-wave caused about 70,000 deaths (Robine et al.,2008). Unfor-tunately, climate change is expected to increase the frequency, duration and intensity of these events (Meehl and Tebaldi, 2004; Dousset et al., 2011; Lemonsu et al., 2013). For Paris, the number of heat-wave days per year on average should increase from one to twelve by the end of the century. Additionally,UHIs also cause a significant increase of cooling demand during summer (Mallick et al.,2009;Stone et al.,2010), and a degradation of air quality by favouring the formation of ozone air pollution and smog (Rosenfeld et al.,1998;Shimoda,2003).
For these reasons, cities are strongly motivated to seek out urban cooling with growing interest from decision-makers. As a result, research efforts can be observed regarding urban cooling techniques. In France, the 2003 heat-wave raised awareness regarding the need to de-ploy adaptation measures for the population, included since 2007 and regularly updated in the Paris Climate Plan (Paris City Hall,2017a). In the scientific literature, various countermeasures have been proposed with that aim, all tackling different mechanisms involved in the urban heat budget. As well, several urban canopy models have been developed and refined over the years, such as the Town Energy Balance (TEB) model (Masson, 2000). Simulations of larger scale implementations provide assistance for decision-makers trying to reduce the impact of
UHIs in cities (Météo-France and CSTB, 2012) and target appropriate UHI-countermeasures based on the urban configuration.
2.2
Urban Heat Budget
The heat budget of an urban surface or volume helps identify the physical parameters driving urban heat flows and that need modification in order to limit the heating of the atmosphere. Figure2.1shows the energy balance of an urban surface.
Urban surface Atmosphere
Convection, H Evaporation, lE
Irradiance, S+L Radiosity, Sup+Lup
Conduction, V
Figure 2.1:Urban surface heat budget.
His the atmospheric convective exchange with the surface, Vis the heat conduction flux,
lEis the latent heat flux if the surface is wet or vegetated (evapotranspiration), withlthe latent heat of vaporization of water andEthe evaporation rate. S andSupare respectively incident and reflected (upward) shortwave radiation (SW), while L and Lup are respectively incident and upward longwave radiation (LW). Radiosity designates the sum of radiation reflected and emitted by the ground (upward flows), while irradiance designates the incident (downward) flows. SWdesignates solar radiation, i.e. both the visible and near-infrared (NIR) bands (0.3 – 3 µm) whileLWcovers the mid-infrared band and part of the far-infrared (3 – 100 µm). The heat budget an urban surface can be summarized by equation2.1:
S+L=Sup+Lup+H+V+lE (2.1)
The radiative balance of incident radiation and radiosity can be condensed into the term
Rn, net radiation, defined as follows:
Rn =S+L− (Sup+Lup) (2.2)
For clear sky conditions in summer, net radiation is positive during the day (radiative ab-sorption) with a predominant component of solar irradiance (S), and negative at night (ra-diative release) with no SW component neither incident nor upward (Oke, 1988b). SWand
LW radiosity can respectively be developed into equations 2.3 and2.4, with αthe albedo of
Tsthe surface temperature. The albedo of a surface represents its solar (SW) reflectivity, while its emissivity quantifies the ability of a surface to emit thermal radiation in proportion to its surface temperature.
Sup= αS (2.3)
Lup= (1−e)L+eσTs4 (2.4)
According toJürges(1924) and to Fourier’s law, respectively, atmospheric convection and heat conduction can be expressed as:
H=h(Ts−Ta) (2.5)
V= −λ∂Tz
∂z |z=0 (2.6)
with h the convective heat transfer coefficient, Ta the air temperature,λ the thermal
con-ductivity of the urban surface and z the depth considered. For its part, the evaporation rate is principally driven by the vapour pressure gradient in the near air, mainly proportional to the water film temperature, but also depending on other meteorological variables such as ra-diation, air pressure, wind speed, etc. Several methods exist to estimate the evaporation rate, either based on the energy or water budget or the mass convection transport problem (MCTP), each making certain assumptions. According toXu and Singh(2001), the latter method offers a satisfying accuracy with limited input data. Derived from theMCTP(Pagliarini and Rainieri,
2011), for a continuous water film whose temperature is that of the surface, the latent flux is:
lE=0.622 lh cap0 Ts ps Ts − pv Ta (2.7)
with ps, pv and p0 respectively the saturation vapour pressure at temperature Ts, partial vapour pressure at temperature Ta and the total air pressure and ca the specific heat of air (1.005 kJ.kg−1.K−1). Therefore, the terms of equation2.1can be developed into the following:
S+L=αS+ (1−e)L+eσTs4+h(Ts−Ta) −λ∂Tz ∂z |z=0 +0.622 lh cap0 Ts ps Ts − pv Ta (2.8)
The left-hand side term of equation2.8represents the inbound flows, while outbound flows are on the right. Among equation2.8’s terms, onlySWandLWreflections (i.e. αS+ (1− e)L)
and irradiance are independent from the surface temperature, which changes over time to balance irradiance with outbound flows.
Figure2.2illustrates the energy balance of an urban volume delimited between the upper border of the urban canopy layer and below a depth into the ground such that the conductive heat flux is negligible over the time scale considered, withQFthe atmospheric anthropogenic
H
H
lE
Rn
Heat storage:
Q
Anthropogenic heat: QF
Q
AFigure 2.2:Urban volume heat budget (Hendel,2020).
heat release,QAthe heat advection outside of the volume and∆Qthe heat storage term in the urban materials. At this scale, heat conductionV is internalized in ∆Q. The energy budget thus becomes:
Rn+QF= H+∆Q+lE+QA (2.9)
For clear skies and low wind speed, it is noteworthy the atmospheric component of the ad-vective termQAis almost null (excluding natural heat advection, e.g. by rivers, or forced heat sink-transfer). Typically, urban areas experience stronger and positive values of atmospheric convectionHboth day and night compared to a rural area, stronger heat storage and anthro-pogenic heat release and less latent flows (Oke,1988b). Regardless of the scale considered,His the term causing atmospheric warming in cities (Asaeda et al.,1996;Christen and Vogt,2004). Existing cooling strategies all aim to reduce this term. This requires the reduction of air and surface temperatures, either by reducing inbound flows or increasing outbound flows, using the following mechanisms:
Reducing inbound flows:
• Reducing irradiance via: − shading (decreaseS)
− lowerLWradiation (decreaseL)
• Reducing atmospheric anthropogenic heat release via: − energy efficiency (decreaseQF)
− heat sink transfer (fromQFtoQA)
Increasing outbound flows:
• Increasing radiosity via:
− higher albedo (increaseSup)
− higher emissivity (increase Lup)
• Increasing the in-depth proportion of conducted/stored heat via: − higher material thermal conductivity (increaseV)
− higher material thermal inertia
− in-depth heat harvesting (increaseVand from∆QtoQA)
• Increasing latent flows (lE) via: − vegetation evapotranspiration − artificial watering
For a cooled surface, convective (H), emittedLWradiation (eσTs4) and conductive flows (V, due to a smaller temperature gradient between the surface and depth) are reduced, resulting in a cooler atmosphere with reduced radiosity. In the literature, the reduction ofLWirradiance is often associated with shading. However, shading devices can defeat this purpose if their temperature rises under solar exposition. It is also noteworthy that most cooling techniques act on several of those physical mechanisms simultaneously. Additional information is provided in Section2.4.
2.3
Pedestrian Heat Stress
Mitigating urban warming mainly addresses the reduction of the air temperature, but also seeks to reduce the other parameters likely to worsen the thermal comfort of a person. Ther-mal comfort is defined byASHRAE(2017) as a "condition of mind that expresses satisfaction with the thermal environment and is assessed by subjective evaluation". Heat stress on the other hand generally refers to a physiological state in which the human body’s thermoregulation is chal-lenged, resulting from an unfavourable heat balance of the human body. The difference be-tween thermal comfort and stress is losely defined, and both terms are often used interchange-ably. However, the word "stress" is meant as less subjective than "comfort", as thermal heat stress depends on the heat budget of the considered individual. The energy balance between the human body and the environment, illustrated in Figure 2.3, depends on several factors: metabolic rate, clothing insulation, air temperature, radiant temperature, air speed and hu-midity.
direct radiation convection clothing reflected radiation infrared radiation infrared radiation
respiration sweat evaporation
external work
conduction
M
Figure 2.3: Human heat budget (adapted fromHavenith,1999). M is the metabolic heat pro-duction.
Several indices for heat stress have been developed over the years, each making certain as-sumptions on the human heat budget. Before presenting some of them, we start by describing the radiative environment, crucial to the pedestrian heat budget.
2.3.1 Radiative Environment and Mean Radiant Temperature
Proper characterization of the radiative environment is crucial to estimating the impact of a
UHI-countermeasure on pedestrian heat stress. Mean radiant temperature (MRTorTmrt) plays a decisive role in thermal comfort in the built environment. It is formally defined as "the tem-perature of a uniform, black enclosure that exchanges the same amount of heat by radiation with the occupant as the actual surroundings" (ASHRAE,2017). In other words, MRTencapsulates inci-dent radiation received in a particular point into one single metric.
A variety of methods exist to estimateMRT, reviewed byGuo et al. (2019). Based on its definition, calculating MRT to the fourth requires summing the radiative flows emitted by each surface at its average temperature, weighed by view factors between the person and the corresponding surface. However, determining all the appropriate view factors is complex, especially for elaborate spaces, even when modelling the human body as a sphere or point.
Experimentally,MRTcan be estimated more easily using different instruments, the most common of which is the globe thermometer. In this case, the simplified geometry of the con-sidered person is that of a sphere. Once the instrument reaches thermal equilibrium, according to Kirchhoff’s law of thermal radiation, the energy balance between convective and radiative heat exchanges can be expressed as follows, with incident flows (ambient radiation) on the left, and outbound flows (black-body emission and convection) on the right:
eσ(Tmrt+273.15)4=eσ(Tg+273.15)4+h(Tg−Ta) (2.10)
withTgthe globe temperature,σthe Stefan-Boltzman constant,ethe globe-thermometer’s
emissivity,Tathe air temperature andhthe convective exchange coefficient, which depends on wind speed and globe diameter. The energy balance of the globe thermometer (equation2.10) can be illustrated with Figure2.4.
h (T
g- T
a)
Figure 2.4: Energy balance of a globe thermometer. Heat gains and losses flows are isotropi-cally received and lost by the globe. Temperatures are expressed in Kelvins.
According toASHRAE(2001), the governing equation from empirical measurements with globe thermometers for estimatingMRTis:
Tmrt= (Tg+273.15)4+ 1.1 ×108v0.6 eD0.4 (Tg−Ta) 0.25 −273.15 (2.11)
with D the globe’s diameter and vthe wind speed. The knowledge of equation2.11’s pa-rameters at a given height thus givesMRT. In principle, this formula applies for any kind of globe thermometer, but the World Meteorological Organization (WMO) standard globe (ISO 7726,1998) is a hollow matte-black copper globe (with emissivity of 0.95), 150 mm in diameter and 0.4 mm in thickness and equipped with a Pt100 resistance thermometer in its center, the measured temperature of which is assumed to be equal to that of the black-globe once thermal equilibrium is reached. This instrument has a typical stabilization time of 20-minutes.
Various globe sizes exist in the literature. Smaller acrylic globes (40 mm in diameter) were found to reach their thermal equilibrium within 5 minutes, and within 4 minutes for acrylic ping-pong balls (Guo et al., 2019; Thorsson et al.,2007;Nikolopoulou et al.,2001). Nonethe-less, smaller globes tend to be more sensitive to wind speed which undermines their accuracy
(Graves,1974). There is general consensus in the literature that there is a trade-off between accuracy of the instrument and its response time, the former and the latter respectively being proportional and inversely proportional to the globe’s diameter (Guo et al.,2019). Furthermore,
Wang and Li(2015) showed that low-conductivity-acrylic globes performed poorly compared
to copper globes under asymmetrical radiation, compromising the hypothesis of isothermal conditions inside the globe.
Modelling tools as well as alternative instruments suited to the estimation ofMRTalso ex-ist, each governed by a specific empirical equation. To cite but a few alternatives, MRTcan be determined with integral radiation measurements (IRM), using a net radiometer simulta-neously measuring short and longwave radiation into the four lateral directions as well as upward and downward (12 instruments in total) (Thorsson et al.,2007). This method is much more demanding than a simple globe thermometer, as it requires computing the mean radiant flux density with the appropriate view factors, but is the most accurate outdoors since it can be adapted to a standing or seated person, when the globe considers equal view factors in all directions instead (Kántor and Unger,2011). Other than that, the constant-air-temperature sen-sor is maintained at the surrounding air temperature with a heat supply, to apply a correction on convective exchanges and deductMRT. Similarly, a two-heating-spheres radiometer can be used to reachMRT. Each sphere has a different emissivity and are maintained at the same tem-perature with a heat supply in order to estimate the convective heat loss to deductMRT(Guo et al.,2019).
Having presented the radiative environment andMRT, we now briefly present some of the most commonly found thermal stress indices in the scientific literature.
2.3.2 Thermal Stress Indices
Researchers have developed several thermal comfort indicators to assess the effects of the envi-ronment on thermophysiology, each making certain assumptions and taking into account more or less input parameters relevant for the human heat budget (Figure2.3). We now present some of the most reported thermal indices in the scientific literature (Binarti et al.,2020).
Wet-Bulb Globe Temperature
One of the most widely used indices is the Wet-Bulb Globe Temperature (WBGT), originally developed to control heat strokes and applied to the United States army (Yaglou et al.,1957). This index is calculated from dry- and wet-bulb (Twet−bulb) air temperatures as well as globe temperature. WBGTthus includes information on radiative environment and air temperature and indirectly on humidity and wind speed through Twet−bulb, but not on metabolic state or clothing. Respectively, for outdoor (with direct shortwave radiation) and indoor environments
WBGTis defined as (Lemke and Kjellstrom,2012):
WBGT =0.7 Twet−bulb+0.2 Tg+0.1 Ta (2.12)
Humidex
The Humidex (HD) is a hot-weather index developped by Canadian meteorologists (
Mas-terton and Richardson, 1979) that empirically scales thermal comfort into bands of thermal
stress, accounting for heat and humidity:
HD= Ta+ 5 9 6.11×exp5417.7530 1 273.16−273.151+Tdew −10 (2.14)
with Tdew the dew point temperature. The exponential factor is a constant based on the latent heat of vaporization of water, the molar mass of water and Boltzmann’s gas constant. Although the Humidex is homogeneous to a temperature, it does not express anHD-equivalent temperature. The correspondence between the index range and thermal stress is given on a scale of comfort. The Humidex equivalent for cold weather is the Wind Chill Index (WCI), calculated on the basis of air temperature and wind speed (Siple and Passel,1945).
Physiological Equivalent Temperature
According to Binarti et al.(2020), the Physiological Equivalent Temperature (PET) is cur-rently the most commonly used outdoor thermal comfort index.PETwas developed to under-stand the impact of meteorological parameters on a "under-standard" person, thus making assump-tions on the metabolic activity and clothing insulation. The index defines an equivalent tem-perature based on the human heat budget under reference conditions. It is the air temtem-perature at which the indoor human heat budget is balanced with the same core and skin temperature as under the actual outdoor conditions (Höppe,1999).
The calculation of PET is based on the Munich Energy Balance Model for Individuals (MEMI), a thermophysiological heat balance model for the human body (Höppe, 1993). The model takes into account meteorological parameters, i.e. air temperature, humidity, wind ve-locity andMRT, and thermophysiological parameters such as heat resistance of clothing (clo units, equivalent to 0.155 K.m².W−1) and human activity (in Watts). TheMEMIincludes basic thermoregulatory processes, such as constriction/dilation of blood vessels and sweat rate. The standard person considered in the model is characterised by a work metabolism of 80 W (light activity) in addition to basic metabolism, and by a clothing heat resistance of 0.9 clo. For the reference indoor setting, MRTis taken equal to air temperature, wind speed is 0.1 m/s, and water vapour pressure is set to 12 hPa (roughly equivalent to a relative humidity (RH) of 50% at a dry air temperature of 20°C) (Matzarakis and Amelung,2008).
ThePET-equivalent temperature is then divided into 9 bands of thermal perception (from "very cold" to "very hot", respectively below -4°C and above 41°C) each corresponding to a level of physiological stress, valid for the standard person considered. Shortcomings ofPET
mainly concern its poor response to clothing insulation and humid conditions. Recently, the
mPET(modified Physiological Equivalent Temperature) was proposed to address this problem and improve thermal comfort predictions (Lin et al.,2019).
Universal Thermal Climate Index
The Universal Thermal Climate Index (UTCI) is a relatively recent index, developped by ex-perts of the International Society of Biometerology with expertise from various fields (thermo-physiology, medicine, physics, biometeorology and environmental sciences). The index takes into account physical variables (air temperature (Ta),MRT(Tmrt), wind speedvand humidity in the form of the saturation vapour pressure, ps) and makes assumptions on the metabolic rate and clothing of pedestrians. As forPET,UTCIis an equivalent temperature defined as the air temperature under reference conditions that would provoke the same thermophysiological response as in the actual environment (Bła ˙zejczyk et al.,2010).
According to Bła ˙zejczyk et al.(2013), hypotheses made on the seasonal clothing and on the metabolic rate are chosen to be representative of most people (average person of 73.5 kg with body fat content of 14% walking at 4 km/h, thus with a metabolic heat production of 135 W/m²). The reference environment is set to be relevant across the broad spectrum of climate zones, and is similar to conditions experienced in an indoor setting, all else being equal. It assumes a wind speed of 0.5 m/s (1.8 km/h) at 10 m height (about 0.3 m/s (1 km/h) at pedes-trian height), aMRT equal to air temperature and a vapour pressure representing a relative humidity of 50% or constant vapor pressure of 20 hPa above 29°C. A clothing model based on seasonal clothing adaptation habits of Europeans is used, adjusted to air temperature and wind speed (Havenith et al.,2012).
The calculation ofUTCIis based on the multinodal Fiala et al. (2012) model, abstracting the human body to twelve spherical and cylindrical body elements built as annular concentric tissue layers. Individual body parts or the whole body can be considered. The model accounts for the complete human heat budget and predicts thermoregulatory reactions such as vasocon-striction and vasodilation of the blood vessels, shivering thermogenesis, and sweat moisture excretion (Bła ˙zejczyk et al.,2013). The fast-calculation script forUTCIwritten byBröde(2009) can be found online freely, for whichUTCIis interpolated with a 6thorder polynomial function
of Ta,v(10 m), psandTmrt−Ta. The model automatically adjusts 10-m-height wind speeds to 1.5 m above ground level. For wind speed measurements made at heightz(different than ten meters), Bröde et al. (2012) provides the following equation based on Oke(1987) to feed the model:
v10m = vz
log(10/0.01)
log(z/0.01) (2.15)
Finally, theUTCIscale is divided into ten levels of thermal stress, for warm or cold condi-tions (from "extreme heat stress" to "extreme cold stress", with UTCI-equivalent temperature respectively ranging from above 46°C to below -40°C). The scale is defined based on the physi-ological response (sweat rate, latent heat loss, skin-core temperature difference, etc.) under the actual environmental conditions.
According to Bła ˙zejczyk et al. (2013), the comparison of UTCI to other thermal comfort indices made from various datasets highlighted that the index was found to better depict bio-thermal conditions for humans than other indices.
Other Indices
Many other heat stress indices exist (De Freitas and Grigorieva, 2015; Teodoreanu, 2016;
Binarti et al.,2020). For example, the Predicted Mean Vote (PMV) empirically fits the human
sensation of thermal comfort based on the prediction of the mean response of a large group of people on the ASHRAE seven-points thermal sensation scale (from -3 "cold" to +3 "hot").
PMVwas later adapted to outdoor settings with the Klima-Michel-Model using weather data as input with assumptions on the activity and clothing (Jendritzky and Nübler,1981). Also, the Standard Equivalent Temperature (SET*), based on theGagge et al.(1986) two-node model, is defined as the equivalent temperature in an environment with a relative humidity of 50%, a wind speed of 0.1 m/s and a MRT equal to air temperature, where a person with standard clothing adapted to its metabolic activity experiences the same heat stress (skin temperature and wetness) as in the actual environment (Binarti et al.,2020).
2.3.3 Cooling Strategies and the Human Heat Budget
Having provided the basics mechanisms behind pedestrian heat stress and indices suited to its assessment, we now briefly discuss implications of urban cooling strategies on the human heat budget.
Cooling strategies most often target urban surfaces, and as such modify their energy bal-ance, by the levers listed in Section2.2, aiming at decreasing inbound and increasing outbound flows. Each term of the urban surface heat budget either has a direct or indirect effect on the parameters impacting the heat balance of a pedestrian. For instance, H will have an influ-ence on air temperature, lEon relative humidity, Sup+Lup impacts the radiative heat budget of the pedestrian, whileV has no significant impact on a standing pedestrian during the day. At night, the heat released by the pavements will tend to increase atmospheric convectionH. This delays the impact of part of the heat absorbed during the day and explains the important magnitude ofUHIat night.
Reducing atmospheric convection thus always has a positive effect both on pedestrians and on the surface. In contrast, other physical mechanisms can have a beneficial cooling effect on the surface, but a negative impact on the pedestrian. For example, evaporation of water on a surface slightly increases the humidity (detrimental to a pedestrian’s heat stress) but signif-icantly reduces its temperature (and thus H, Sup, Lup and V, all of which is beneficial for a pedestrian). Similarly, increasing the albedo of a surface reduces its temperature (and thusH,
Lup and V), but by definition increases the SW radiosity Sup, unfavourable for the radiative budget of a person.
In such cases, care must be given to ensuring that the potential drawbacks of a countermea-sure do not offset the benefits on the pedestrian’s heat balance. In practice, such a condition is rarely achieved, except for very specific cases, depending for example on the urban configura-tion. This is further discussed in the following Section (2.4), describing urban heat mitigation strategies.
